Darwin's Black Box: The Biochemical Challenge to Evolution

work. Perhaps by this point in the book, the reader can easily see how the transport system that sent the battery crusher to its destination is irreducibly complex. If any of its numerous components is missing, then the crusher is not delivered to the garbage treatment room. Furthermore, the delicate balance of the system must be maintained; each of the many components that interlock must do so precisely and then disengage, and each must arrive and depart at the proper times. Any single error will cause the system to fail. REALITY CHECK This is science fiction, isn’t it? Things this complex don’t exist in nature, do they? The cell is a “homogeneous globule of protoplasm,” isn’t it? Well, no, yes, and no. All of the fantastic machines in our space probe have direct counterparts in the cell. The space

probe itself is the cell, the library is the nucleus, the blueprint is the DNA, the copy of the blueprint is RNA, the window of the library is the nuclear pore, the master machines are ribosomes, the main area is the cytoplasm, the ornament is the signal sequence, the battery crusher is a lysosomal hydrolase, the guide is the signal recognition particle (SRP), the receiving site is the SRP receptor, processing room 1 is the endoplasmic reticulum (ER), processing rooms 2 through 4 are the Golgi apparatus, the antenna is a complex carbohydrate, the sub-rooms are coatomer or clathrin-coated vesicles, and various proteins play the roles of the trimmer, hauler, delivery coder, port marker, and gateway. The garbage treatment room is the lysosome. Let’s quickly run through a description of how a protein that is synthesized in the cytoplasm eventually finds its way to the lysosme. This will take just one paragraph. Don’t worry if you rapidly forget the names and procedures of cellular

transport; the purpose is simply to give you a glimpse of the cell’s complexity. An RNA copy (called messenger RNA, or just mRNA) is made of the DNA gene coding for a protein that works in the cell’s garbage disposal—the lysosome. We’ll call the protein “garbagease.” The mRNA is made in the nucleus, then floats over to a nuclear pore. Proteins in the pore recognize a signal on the mRNA, the pore opens, and the mRNA floats into the cytoplasm. In the cytoplasm the cell’s “master machines”—ribosomes—begin making garbagease using the information in the mRNA. The first part of the growing protein chain contains a signal sequence made of amino acids. As soon as the signal sequence forms, a signal recognition particle (SRP) grabs onto the signal and causes the ribosome to pause. The SRP

and associated molecules then float over to an SRP receptor in the membrane of the endoplasmic reticulum (ER) and stick there. This simultaneously causes the ribosome to resume synthesis and a protein channel to open in the membrane. As the protein passes through the channel and into the ER, an enzyme clips off the signal sequence. Once in the ER, garbagease has a large, complex carbohydrate placed on it. Coatomer proteins cause a drop of the ER, containing some garbagease plus other proteins, to pinch off, cross over to the Golgi apparatus, and fuse with it. Some of the proteins are returned to the ER if they contain the proper signal. This happens two more times as the protein progresses through the several compartments of the Golgi. Within the Golgi an enzyme recognizes the signal patch on garbagease and places another carbohydrate group on

it. A second enzyme trims the freshly attached carbohydrate, leaving behind mannose-6-phosphate (M6P). In the final compartment of the Golgi, clathrin proteins gather in a patch and begin to bud. Within the clathrin vesicle is a receptor protein that binds to M6P. The M6P receptor grabs onto the M6P of garbagease and pulls it on board before the vesicle buds off. On the outside of the vesicle is a v-SNARE protein that specifically recognizes a t-SNARE on the lysosome. Once docked, NSF and SNAP proteins fuse the vesicle to the lysosome. Garbagease has now arrived at its destination and can begin the job for which it was made. The fictional space probe is so complicated it hasn’t been invented yet, even in a crude way. The authentic cellular system is already in place, and

every second of every day, this process happens uncounted billions of time in your body. Science is stranger than fiction. THE DEMANDS OF THE JOB Garbagease travels a distance of about one ten- thousandth of an inch on its journey from the cytoplasm to the lysosome, yet it requires the services of dozens of different proteins to ensure its safe arrival. In our imaginary TV movie, the vaccine traveled perhaps a thousand miles from the Centers for Disease Control to the big city where it was needed—a trillion times farther than garbagease traveled. But many of the requirements for transporting the vaccine were the same as those for getting the enzyme from the cytoplasm to the lysosome. The demands are imposed by the type of task to be done; they don’t depend on the distance traveled, the type of vehicle used, or the materials out of which the signs are made.

A current textbook distinguishes three methods that the cell uses to get proteins into 1 compartments. The first, where a large gate opens or closes to regulate the passage of proteins through the membrane, is known as gated transport. This is the mechanism that regulates the flow of material such as newly-made mRNA between the nucleus and the cytoplasm (or in space-probe language, the flow of the blueprint out of the library into the main area). The second method is transmembrane transport. This occurs when a single protein is threaded through a protein channel, as when garbagease passed from the cytoplasm into the ER. The third way is vesicular transport, where protein cargo is loaded into containers for shipment, as happened for the trip from the Golgi (the final processing room) to the lysosome (the garbage treatment room). For our purposes the first two methods can be considered to be the same: they both use portals in a membrane that selectively allow proteins

through. In the case of gated transport the portal is quite large, and proteins can pass through in their folded form. In the case of transmembrane transport the portal is smaller, and proteins must be threaded through. But in principle there is no roadblock to expanding or contracting the size of a portal, so these are equivalent. Therefore I will call both of these gated transport. What are the bare, essential requirements for gated transport? Imagine a parking garage that is reserved for persons with diplomatic license plates. In place of a human attendant the garage has a scanner that reads a barcode on the license plate, and if the barcode is correct the garage door opens. A car with diplomatic plates drives up, the scanner scans the barcode, the door opens, and the car drives in. It doesn’t matter if the car drove ten feet to the garage or ten thousand miles, or whether the vehicle is a truck, jeep, or motorcycle; if the barcode is there, it can pass through. Thus three basic components are required for gated

transport at the garage: an identification tag; a scanner; and a gate that is activated by the scanner. If any of these things are missing, then either the vehicle does not get in or the garage is no longer a reserved area. Because gated transport requires a minimum of three separate components to function, it is irreducibly complex. And for this reason the putative gradual, Darwinian evolution of gated transport in the cell faces massive problems. If proteins contained no signal for transport, they would not be recognized. If there were no receptor to recognize a signal or no channel to pass through, again transport would not take place. And if the channel were open for all proteins, then the enclosed compartment would not be any different from the rest of the cell. Vesicular transport is even more complicated than gated transport. Suppose now that, instead of the diplomats’ cars entering the garage one at a time, all diplomats had to drive their cars into the back

of a large tractor-trailer truck, the truck would drive into the special garage, and the cars would drive off the truck and park. Now we need a way for the truck to recognize the proper cars, a way for the garage to recognize the truck, and a way for the cars to get out of the truck inside the garage. Such a scenario requires six separate components: (1) an identification tag on the cars; (2) a truck that can carry the cars; (3) a scanner on the truck; (4) an identification tag on the truck; (5) a scanner on the garage; (6) an activatable garage gate. In the cell’s vesicular transport system these components correspond to mannose-6-phosphate, the clathrin vesicle, the M6P receptor in the clathrin vesicle, v-SNARE, t-SNARE, and SNAP/NSF proteins. In the absence of any of these functions, either vesicular transport cannot take place or the integrity of the destination compartment is compromised. Because vesicular transport requires several more components than gated transport, it cannot

develop gradually from gated transport. For example, if we had barcode stickers on the diplomats’ cars, placing cars inside a truck (a vesicle to transport them) would hide the stickers, and they would fail to enter the garage. Or suppose instead that the truck had the same label that the cars had, so it could enter the garage. But we would still be missing a mechanism to get the cars on the truck, so the truck would be of no use. If some cars randomly entered the truck then, again, nondiplomats’ cars would enter the garage. Returning to the world of the cell, if a vesicle just “happened” to form there would be no mechanism for identifying the proteins that should enter it, and no way to specify its destination. Placing proteins containing address labels into an unlabeled vesicle would make the labels unavailable, and therefore would be detrimental to the organism that had a happily functioning gated transport system. Gated transport and vesicular transport are two separate mechanisms; neither

helps in understanding the other. The brief sketch of the requirements for gated and vesicular transport in this chapter did not take into account many complexities of the systems. But since these only make the system more intricate, they cannot ameliorate the irreducible complexity of targeted transport. SECOND-HAND ROSE Irreducibly complex systems like mousetraps, Rube Goldberg machines, and the intracellular transport system cannot evolve in a Darwinian fashion. You can’t start with a platform, catch a few mice, add a spring, catch a few more mice, add a hammer, catch a few more mice, and so on: The whole system has to be put together at once or the mice get away. Similarly, you can’t start with a signal sequence and have a protein go a little way towards the lysosome, add a signal receptor protein, go a little further, and so forth. It’s all or

nothing. Perhaps, though, we’re overlooking something. Perhaps one of the parts of a mousetrap was used for some purpose other than trapping mice, and so were the other parts. At some point several parts that were being used for other purposes suddenly came together to produce a functioning trap. And perhaps the components of the intracellular transport system were originally performing other tasks in the cell, then switched to their present role. Could that happen? An exhaustive consideration of all possible roles for a particular component can’t be done. We can, however, consider a few likely roles for some of the components of the transport system. Doing so shows it is extremely implausible that components used for other purposes fortuitously adapted to new roles in a complex system. Suppose we start with a protein that because it had an oily region, resided in the cell’s membrane. Suppose further that it was beneficial for the

protein to be there because it toughened the membrane, making it resistant to tears and holes. Could that protein somehow turn into a gated channel? This is like asking if wooden beams in a wall could be transformed, step by Darwinian step, small mutation by small mutation, into a door with a scanner. Suppose wooden beams were brought together, and the area between them was weakened so much that plaster cracked and a hole formed in the wall. Would that be an improvement? The hole in the wall would let insects, mice, snakes, and other things into the room; it would let heat or air-conditioning out. Similarly, a mutation that caused proteins to aggregate in the membrane, leaving a small hole, would let stored foodstuffs, salt, ATP, and other needed materials float away. That is no improvement. A house with a hole in the wall would never sell, and a cell with a hole in it would be at a great disadvantage compared to other cells.

Suppose instead that a protein could bind to the beginnings of new proteins as they were being put together by the ribosome. Suppose that was an improvement because new, unfolded proteins are more vulnerable, so placing a folded protein on them would protect them until they were fully made and folded. Could such a protein develop into, say, the signal recognition particle (SRP)? No. Such a protein would help a new protein fold rapidly, not keep it unfolded—the opposite of what modern SRP does. Folded proteins, however, can’t get through the gated channel where the modern SRP takes them. Further, if a proto-SRP caused the ribosome to halt its synthesizing, as the modern SRP does, but the machinery to turn the ribosome back on was not yet in place, then that would kill a cell (some deadly poisons kill by

turning off the cell’s ribosomes). So we have a dilemma: in the beginning an uncontrolled inhibitor of protein synthesis would kill the cell, but a temporary halt in protein synthesis is crucial in modern cells. If the ribosome does not pause, the new protein gets so big that it can’t fit through a gated channel. So it appears that the modern SRP could not have developed from a protein whose job it was to bind new proteins and protect them from degradation. Suppose that an enzyme placed a large carbohydrate group (the “bauble”) on proteins as they were made. Suppose that helped stabilize the protein somehow, making it last longer in the cell. Could that step eventually become part of the intracellular transport chain? No. The bauble, because it would make the protein larger, would prevent it from passing through any future gate

that looked like a modern gate in the ER. The bauble would actually be a hindrance to developing a transport system. In the same way, other isolated parts of the system would actually be damaging to the cell, not helpful. An enzyme that clipped off the signal sequence (the “ornament”) would be detrimental if the signal sequence was playing a positive role in a primitive cell. Trimming of the bauble would be a step backward if the bauble had a job to do. Trapping of proteins like “garbagease” inside a vesicle would be harmful if garbagease originally had to work in the open. In Chapter 2 I noted that one couldn’t take specialized parts of other complex systems (such as the spring from a grandfather clock) and use them directly as specialized parts of a second irreducible system (like a mousetrap) unless the parts were first extensively modified. Analogous parts playing other roles in other systems cannot

relieve the irreducible complexity of a new system; the focus simply shifts from “making” the components to “modifying” them. In either case, there is no new function unless an intelligent agent guides the setup. In this chapter we see that construction of a transport system faces the same problem: the system can’t be put together piecemeal from either new or secondhand parts. DEATH AT AN EARLY AGE In one version of our made-for-TV movie, a wrong label was placed on a carton of vaccine, and children died. Fortunately, it was only make- believe: a story about a story. But in real life, mixed-up or missing labels can cause real deaths. A crying two-year-old girl stands in front of a height chart, with the aid of an adult’s helping hand. She is only two feet tall. Her face and eyes are puffed up, and her legs are bent. She moves stiffly. She is severely retarded. A medical

examination shows an enlarged heart, liver, and spleen. A cough and runny nose bespeak another of the many upper respiratory infections she has endured in her young life. The doctor takes a tissue sample from the girl and sends it to a lab for analysis; a lab worker grows cells from the sample in a Petri dish and examines them under a microscope. Each of the cells contains thousands of little, dense grains that aren’t present in normal cells. The grains are called “inclusion bodies”; the 2 little girl has I-cell disease. Because the disease is progressive, the skeletal and neural difficulties will increase with time. The girl will die before the age of five. I-cell disease is caused by a defect in the protein transport pathway. The cells of patients with the disease lack one of the machines in the long chain that takes proteins from the cytoplasm to the lysosome. Because of the defect, enzymes intended for the lysosome never make it there. Instead they are shunted off in the wrong vesicle to the cell

membrane and dumped into the extracellular space. The cell is a dynamic system, and just as it must build new structures, it must continually degrade old ones. Old material is brought to the lysosome for degradation. In children with I-cell disease, the garbage is dumped into the disposal as it should be, but the disposal is broken: neither “garbagease” nor any other degradative enzyme that normally decomposes old structures is present. As a result garbage piles up, and lysosomes get filled. The cell makes new lysosomes to accomodate the increasing waste, but the new compartments eventually fill up with the detritus of cellular life. Over time the entire cell becomes bloated, tissues become enlarged, and the patient dies. A child can die because of this single defect in one of the many machines needed for taking proteins to the lysosome. A single flaw in the cell’s labyrinthine protein-transport pathway is fatal.

Unless the entire system were immediately in place, our ancestors would have suffered a similar fate. Attempts at a gradual evolution of the protein transport system are a recipe for extinction. Because of the medical problems associated with the failure of the transport system, and because the system is so intricate and fascinating, we might expect the evolutionary development of vesicular protein transport to be a busy area of research. How could such a system develop step-by-step? What hurdles would the cell have to overcome as it moved from some other method of dealing with garbage to a coated vesicle specifically targeted to, and equipped for merger with, the lysosome? Once again, if we looked in the literature for an explanation of the evolution of vesicular transport, we would be crushingly disappointed. Nothing is there. Annual Review of Biochemistry (or ARB)is a book series, very popular with biochemists, that reviews the current state of knowledge in selected research

areas. In 1992 an article was published in ARB concerning “Vesicle-Mediated Protein Sorting.” 3 The authors begin their review by stating the obvious: “The transport of proteins between membrane-bounded organelles is an immensely complex process.” They proceed in professional fashion to describe the systems and current research in the area. But we can read from one end of the forty-six-page review to the other without encountering an explanation for how such a system might have gradually evolved. The topic is off the radar screen. Logging on to a computer database of the professional literature in the biomedical sciences allows you to do a quick search for key words in the titles of literally hundreds of thousands of papers. A search to see what titles have both evolution and vesicle in them comes up completely empty. Slogging through the literature the old-fashioned way turns up a few scattered papers that speculate on how gated transport

between compartments of a eukaryotic cell might 4 have developed. But all the papers assume that the transport systems came from preexisting bacterial transport systems that already had all the components that modern cells have. This does us no good. Although the speculations may have something to do with how transport systems could be duplicated, they have nothing to do with how the initial systems got there. At some point this complex machine had to come into existence, and it could not have done so in step-by-step fashion. Perhaps the best place to get an overview of vesicle transport is from the textbook Molecular Biology of the Cell by National Academy of Science President Bruce Alberts, Nobel Prize winner James Watson, and several more coauthors. The textbook spends 100 pages on the elegant details of gated and vesicular transport. 5 In that 100 pages there is a one-and-a-half-page section entitled “The Topological Relationships of Membrane-Bounded Organelles Can Be

Interpreted in Terms of Their Evolutionary Origins.” In this section the authors point out that if a vesicle pinches off from the cell membrane and into the cell, then its inside is equivalent to the outside of the cell. They then suggest that the nuclear membrane, ER, Golgi, and lysosomes first arose when parts of the cell membrane pinched off. This may or may not be true, but it does not even address the origin of protein transport, either vesicular or gated. Clathrin is not mentioned in this short section, nor are the problems of loading the correct cargo into the correct vesicle and targeting it to the correct compartment. In short, the discussion is irrelevant to the questions we are asking. At the end of our literature search, we know no more than when we started. SUMMING UP AND LOOKING AHEAD Vesicular transport is a mind-boggling process, no less complex than the completely automated delivery of vaccine from a storage area to a clinic a

thousand miles away. Defects in vesicular transport can have the same deadly consequences as the failure to deliver a needed vaccine to a disease-racked city. An analysis shows that vesicular transport is irreducibly complex, and so its development staunchly resists gradualistic explanations, as Darwinian evolution would have it. A search of the professional biochemical literature and textbooks shows that no one has ever proposed a detailed route by which such a system could have come to be. In the face of the enormous complexity of vesicular transport, Darwinian theory is mute. In the next chapter I will examine the art of self- defense—but, of course, on a molecular scale. Just as machine guns, battle cruisers, and nuclear bombs are necessarily sophisticated machines in our larger world, we will see that tiny cellular defense mechanisms are quite complex, too. Few things are simple in Darwin’s black box.

CHAPTER 6 ALL SHAPES AND SIZES Enemies abound. Paranoia has nothing to do with it; we are surrounded by creatures that, for one reason or another, want to do us in. Since most people don’t want to die just yet, they take steps to defend themselves. Threats of aggression can come in all shapes and sizes, so defenses have to be versatile. The largest- scale threat is war between nations. Rulers of nations always seem to be wanting the resources of neighboring countries, so threatened countries have to defend themselves or suffer unpleasant consequences. In modern times, countries can have very sophisticated means of defense indeed. The United States has stockpiled atomic bombs; if

some other country shakes its proverbial fist at us, we can rattle our bombs at them. If threats escalate to violence and we don’t wish to use atomic bombs for one reason or another, then other machines can be deployed: jets that drop “smart” bombs, AWACS planes that monitor the air space for many miles, tanks equipped for night combat, surface-to-air-missiles that shoot down surface-to-surface missiles, and much more. To the techno-war-monger, we live in a golden age. Big threats like war are important, but other types of aggression can kill, too. Terrorist bombings of planes or gas attacks on subways have, unfortunately, become too frequent for comfort. Worse, none of the weapons mentioned above will help much to prevent a subway gas attack. When the nature of the enemy changes dramatically— from a foreign country to a domestic terrorist group—the nature of the defense must also change. Instead of bombs, government officials install metal detectors at airports and place guards

with guns at strategic locations. Terrorism and war threaten us, but they happen infrequently. On a day-to-day basis more people are assaulted by muggers and mayhem in their neighborhood than by exotic groups or foreign countries. The streetwise city dweller will have bars on his window, use an intercom or peephole to see who is at the door, and carry a can of pepper spray when it’s time to walk the dog. In lands where such modern conveniences are unknown, stone or wooden walls can be built around the hut to keep out intruders (both two-and four-footed), and a spear is kept by the bed in case the wall is breached. A stick, rock, barrier, gun, alarm, tank, and atomic bomb can all be used to help fend off attacks. Since the circumstances in which each weapon is useful might vary considerably, there is a lot of overlap. Both a stick and a pistol can deter a mugger; a pistol and a tank can threaten a terrorist group; and both a tank and an atomic bomb can be

used against a foreign country. Looked at this way, we can speak about the “evolution” of defensive systems. We can talk about an arms race in which the equipment of competing sides becomes more and more sophisticated. We can tell stories about life being a struggle where people or countries with the best defenses survive. But before we hop in a box and fly off with Calvin and Hobbes, we need to recall the distinction between conceptual precursors and physical precursors. A rock and a gun can both be used for defense, but a rock cannot be turned into a gun by a series of small steps. A can of pepper spray is not a physical precursor of a hand grenade. A jet plane cannot be changed into an atomic bomb one nut and bolt at a time, even though both the plane and the bomb do contain nuts and bolts. In Darwinian evolution, only physical precursors count. Humans and large animals are not the only threats a person encounters. There are also Lilliputian aggressors against whom bombs or guns or rocks

are ineffective. Bacteria, viruses, fungi—they all would xlove to eat us if they could. Sometimes they do, but most times they don’t because our bodies have an array of defensive systems to deal with microscopic attacks. The first line of defense is the skin. Like a stockade fence, the skin works by a relatively low-tech method: it’s a barrier that is hard to breach. Burn victims often succumb to massive infections because the skin barrier has been broken and the internal defenses can’t cope with the overwhelming numbers of invaders. But although skin is an important part of the body’s defense, it is not a physical precursor of the immune system. To discourage any outsider who manages to climb to the top, sometimes stockade walls have spikes on them. Where I lived in the Bronx, almost all of the cyclone fences were topped with razor wire, which apparently is more effective at lacerating intruders than old-fashioned barbed wire. Spikes and razor wire are not parts of the fence proper;

they are little add-ons that increase the effectiveness of the barrier. Still, like the fence itself, razor wire is not a physical precursor to, say, a gun or a landmine. Skin, too, has add-ons that increase its effectiveness as a barrier. In a biochemistry laboratory you often have to wear gloves to protect yourself from the material you’re handling, but sometimes you have to wear gloves to protect the material from you. People who work with RNA wear gloves because human skin excretes an enzyme that chops up RNA. Why? It turns out that many viruses are made from RNA. To such a virus, the enzyme is like razor wire on the skin: any RNA that tries to breach the barrier gets lacerated. There are other types of spikes on the skin. One of the most interesting is a class of molecules called magainins, discovered by a biologist named Mike Zasloff after he wondered why live laboratory frogs that are cut open and sewed back up in

nonsterile conditions rarely get infections. He showed that their skin excretes a substance which can kill bacterial cells; since then, magainins have been discovered in many kinds of animals. But magainins, like the RNA-destroying enzymes, are not precursors to the sophisticated defense systems under the skin of animals. To find the heavy weaponry, we have to peek under our skins. The internal defense system of vertebrates is dizzyingly complicated. Like the modern U.S. army, it has a variety of different weapons that can overlap in their use. But like the weapons we discussed above, we must not automatically assume the different parts of the immune system are physical precursors of each other. Although the body’s defenses are still an active area of research, much is known in detail about particular aspects. In this chapter I will discuss selected parts of the immune system and point out the problems they present for a model of gradual evolution. Those who become intrigued by

the cleverness of the systems and want to know more are encouraged to pick up any immunology text for the details. 1 THE RIGHT STUFF When a microscopic invader breaches the outer defenses of the body, the immune system swings into action. This happens automatically. The molecular systems of the body, like the Star Wars anti-missile system that the military once planned, are robots designed to run on autopilot. Since the defense is automated, every step has to be accounted for by some mechanism. The first problem that the automated defense system has is how to recognize an invader. Bacterial cells have to be distinguished from blood cells; viruses have to be distinguished from connective tissue. Unlike us, the immune system can’t see, so it has to rely initially on something akin to a sense of touch. Antibodies are the “fingers” of the

blind immune system—they allow it to distinguish a foreign invader from the body itself. Antibodies are formed by an aggregation of four chains of amino acids (Figure 6-1): two identical light chains, and two identical heavy chains. The heavy chains are about twice as big as the light chains. In the cell, the four chains make a complex that resembles the letter Y. Because the two heavy chains are the same and the two light chains are the same, the Y is symmetrical: if you took a knife and cut it down the middle you’d get identical halves, with one heavy and one light chain in each half. At the end of each pronged tip of the Y there is a depression (called a binding site). Lining the binding site are portions of both the light chain and the heavy chain. Binding sites come in a large variety of shapes. One antibody might have a binding site with a piece jutting up

here, a hole over there, and an oily patch on the edge. A second antibody might have a positive charge on the left, a crevice in the middle, and a bump on the right. FIGURE 6-1 SCHEMATIC DRAWING OF AN ANTIBODY MOLECULE. If the shape of a binding site just happens to be exactly complementary to the shape of a molecule on the surface of an invading virus or bacterium, then the antibody will bind to that molecule. To get a feel for it, imagine a household object with a depression in it and a few knobs poking up out of the depression. My youngest daughter has a doll wagon with front and back seats—something like

that will do nicely. Now take the wagon/object, go around the house, and see how many other articles will fit snugly into the depression, filling both the front seat and the back seat without leaving any spaces. If you find even one, you’re luckier than I am. Nothing in my house fit snugly in the wagon, and neither did anything in my office or laboratory. I imagine there’s some object out in the world with a shape complementary to the wagon’s, but I haven’t found it yet. The body has a similar problem: the odds of any given antibody binding to any given invader are pretty slim. To make sure that at least one kind of antibody is available for each attacker, we make billions to trillions of them. Usually, for any particular invader, it takes 100,000 to find one antibody that works. When bacteria invade the body, they multiply. By the time an antibody binds to a bacterium there may be many, many copies of the bug floating around. Against this Trojan horse that breeds, the

body has 100,000 guns, but only one works. One handgun isn’t going to do much good against a horde; somehow reinforcements have to be brought in. There’s a way to do this, but first I have to back up and explain a bit more about where antibodies come from. There are billions of different kinds of antibodies. Each kind of antibody is made in a separate cell. The cells that make antibodies are called B cells, which is easy to remember because they are 2 produced in the bone marrow. When a B cell is first born, mechanisms inside of it randomly choose one of the many antibody genes that are encoded in its DNA. That gene is said to be “turned on”; all other antibody genes are “turned off.” So the cell produces only one kind of antibody, with one kind of binding site. The next cell that’s made will in all likelihood have a different antibody gene turned on, so it will make a different protein with a different binding site. The principle, then, is one cell, one type of

antibody. Once a cell commits to making its antibody, you might think that the antibody would leave the cell so it could patrol the body. But if the contents of all B cells were dumped out into the body, there would be no way to tell which cell the antibody came from. The cell is the factory that makes the particular type of antibody; if the antibody finds a bacterium, we need to tell the cell to send us reinforcements. But with this hypothetical setup, we can’t get a message back. Fortunately, the body is smarter than that. When a B cell first makes its antibody, the antibody anchors in the cell membrane with the prongs of the Y sticking out (Figure 6-2). The cell does this trick by using the gene for the normal antibody, and also using a little piece of a gene that codes for an oily tail on the protein. Since the membrane is oily, too, the piece sticks in the membrane. This step is critical, because now the binding site of the antibody is attached to its factory. The entire B cell

factory patrols the body; when a foreign invader enters, the antibody-with-attached-cell binds. Now we have the factory close at hand to the invaders. If the cell could be signaled to make more of the antibody, then the fight would be helped by reinforcements. Fortunately, there is a way to send a signal; unfortunately, it’s pretty convoluted. When an antibody on a B cell binds to a foreign molecule it triggers a complex mechanism to swallow the invader: in effect, the munitions factory takes a hostage. The antibody then breaks off a piece of membrane to make a little vesicle—a self-made taxicab. In this taxi, the hostage is brought into the B-cell factory. Inside the cell (still in the cab) the foreign protein is chopped up, and a piece of the foreign protein sticks to another protein (called an MHC protein). The cab then returns to the membrane of the cell. Outside the factory, along comes another cell (called a helper T cell). The helper T cell binds to the B cell, which is “presenting” the chopped-up

piece of invader (the foreign fragment in the MHC protein) for the T cell’s consideration. If the fit is just right, it causes the helper T cell to secrete a substance called interleukin. Interleukin is like a message from the Department of Defense to the munitions factory. By binding to another protein on the surface of the B cell, the interleukin sets off a chain of events that sends a message to the nucleus of the B cell. The message is: grow! FIGURE 6-2 SCHEMATIC DRAWING OF A BCELL. The B cell begins to reproduce at a rapid rate. T cells continue to secrete interleukin if they are

bound to a B cell. Eventually the growing B-cell factory produces a series of spinoff factories in the form of specialized cells called ‘plasma cells.’ Instead of producing a form of the antibody that sticks in the membrane, plasma cells leave off the last oily piece of the protein. Now free antibody is extruded in large amounts into the extracellular fluid. The switch is critical. If the new plasma-cell factories were like the old B-cell factory, the antibodies would all be confined to quarters and would be much less effective at inhibiting the invaders. STEP BY STEP Could this system have evolved step-by-step? Consider the vast pool of billions to trillions of factory B cells. The process of picking the right cell out of a mixture of antibody-producing cells is called clonal selection. Clonal selection is an elegant way to mount a specific response in great numbers to a wide variety of possible foreign

invaders. The process depends on a large number of steps, some of which I have not discussed yet. Leaving those aside for now, let’s ask what the minimum requirements are for a clonal selection system, and if those minimum requirements could be produced step-by-step. The key to the system is the physical connection of the binding ability of the protein with the genetic information for the protein. Theoretically this could be accomplished by making an antibody where the tail of the Y bound to the DNA that coded for the protein. In real life, however, such a setup wouldn’t work. The protein might be connected to its genetic information, but because the cell is surrounded by a membrane, the antibody would never come in contact with the foreign material, which is floating around outside the cell. A system where both the antibody and its attached gene were exported from the cell would overcome that problem, only to run into a different one: outside the cell there would be no cellular

machinery to translate the DNA message into more protein. Anchoring the antibody in the membrane is a good solution to the problem; now the antibody can mix it up with a foreign cell and still be near its DNA. But although the antibody can bind the foreign material without floating away from the cell, it does not have direct physical contact with the DNA. Since the protein and DNA are blind, there must be a way to get a message from one to the other. Just for now, for the sake of argument, let’s forget about the tortuous way that the message of binding actually gets to the B-cell nucleus (requiring the taxicab, ingestion, MHC, helper T cells, interleukin, and so on). Instead let’s imagine a simpler system where there’s only one other protein. Let’s say that when the antibody binds to a foreign molecule, something happens that attracts some other protein—a messenger to take word of a hostage to the factory nucleus. Maybe

when the hostage is first found, the shape of the antibody changes, perhaps pulling up a little on the antibody’s tail. Perhaps part of the antibody’s tail sticks into the inside of the cell, which is what triggers the messenger protein. The change in the tail could cause the messenger protein to scuttle into the nucleus and bind to the DNA at a particular point. Binding to the right place on the DNA is what causes the cell to start growing and to start producing antibody without the oily tail— antibody that gets sent out of the cell to fight the invasion. Even in such a simplified scheme, we are left with three critical ingredients: (1) the membrane-bound form of the antibody; (2) the messenger; and (3) the exported form of the antibody. If any of these components is missing, the system fails to function. If there is no antibody in the membrane, then there’s no way to connect a successful antibody that binds a foreign invader to the cell containing the genetic information. If there is no

exported form of the antibody, then when the signal is received there is nothing to send out into the world to fight. If there is no messenger protein, then there is no connection between binding the membrane antibody and turning on the right gene (making the system about as useful as a doorbell whose wires had been cut). A cell hopefully trying to evolve such a system in gradual Darwinian steps would be in a quandary. What should it do first? Secreting a little bit of antibody into the great outdoors is a waste of resources if there’s no way to tell if it’s doing any good. Ditto for making a membrane-bound antibody. And why make a messenger protein first if there is nobody to give it a message, and nobody to receive the message if it did get one? We are led inexorably to the conclusion that even this greatly simplified clonal selection could not have come about in gradual steps. Even at this simplified level, then, all three ingredients had to evolve simultaneously. Each of

these three items—the fixed antibody, the messenger protein, and the loose antibodies—had to be produced by a separate historical event, perhaps by a coordinated series of mutations changing preexisting proteins that were doing other chores into the components of the antibody system. Darwin’s small steps have become a series of wildly unlikely leaps. Yet our analysis overlooked many complexities: How does the cell switch from putting the extra oily piece on the membrane to not putting it on? The message system is fantastically more complicated then our simplified version. Ingestion of the protein, chopping it up, presenting it to the outside on an MHC protein, specific recognition of the MHC/fragment by a helper T cell, secretion of interleukin, binding of interleukin to the B cell, sending the signal that interleukin has bound into the nucleus—the prospect of devising a step-by- step pathway for the origin of the system is enough to make strong men blanch.

MIX AND MATCH Factories float around in huge numbers, poised to deliver antibodies that can stick to an invader with virtually any shape. But how does the body make all those billions of differently shaped antibodies? It turns out that there is an elegant trick for making very many different antibodies without requiring enormous quantities of genetic material to code for the proteins. Over the next few pages I’ll describe the system in some detail. Again, don’t be concerned if the details quickly slip your mind; my purpose here is just to help you appreciate the complexity of the immune system. It took a fascinating discovery to lead scientists to puzzle out the full complexity of the immune system. The discovery started with a potentially cruel, but necessary, experiment. Just to see what would happen, chemists made some small molecules that do not occur in nature and

then attached them to a protein. When the protein carrying the synthetic molecules was injected into a rabbit, the scientists were astonished to find that, yes, the rabbit made antibodies that bound tightly to the synthetic molecule. How could this be? Neither the rabbit nor its ancestors ever met the synthetic molecule, so how did it know how to make antibodies against it? Why should it recognize a molecule it had never seen before? The puzzle of “antibody diversity” intrigued scientists studying immunology. Several ideas were floated as possible explanations. Proteins were known to be flexible molecules, and antibodies are proteins. So maybe when a new molecule is injected into the body an antibody wraps around it, molds itself to that shape, and then somehow freezes in that configuration. Or maybe, because defense is so vitally important, the

DNA of organisms contains a vast number of genes for antibodies with many different shapes— enough to allow them to recognize things they hadn’t seen yet. But such a huge number of antibodies would take up more than the available coding space in the DNA. So maybe there were only a few antibodies, and when the cell divided, maybe there was some way to make a lot of mutations in just the areas coding for the binding sites of the antibodies. That way each new B cell in the body could carry different mutations, coding for an antibody different from all other B cells. Or maybe the answer was a combination of these, or maybe it involved something completely new. The answer to the problem of antibody diversity had to await an astonishing discovery: a gene coding for a protein didn’t always have to be a continuous segment of DNA—it could be 3 interrupted. If we compare a gene to a sentence, it was as if a protein’s code, “The quick brown fox jumps over the lazy dog” could be altered (without

destroying the protein) to read “The quick brdkdjf bufjwkw nhruown fox jumps over the lapfeqmzda lfybnek sybagjufu zy dog.” The sensible DNA message was broken up by tracts of nonsense letters that somehow were not included in the protein. Further work showed that for most genes, corrections would be made—splicing out the nonsense—after an RNA copy is made of a DNA gene. Even with “interrupted” DNA, an edited and corrected message in RNA could be used by the cell’s machinery to make the correct protein. Even more surprisingly, for antibody genes the DNA itselfcan also be spliced. In other words, DNA that is inherited can be altered. Amazing! Splicing and rearrangement of DNA play a large role in explaining the great number of antibodies that the body can produce. The following is a brief description of work that has taken many investigators many years to accomplish; because of their efforts, the riddle of antibody diversity is solved.